DESCRIPTION (Applicant's Abstract): It is now accepted that pregnancy and birth complications (PBC's) are associated with increased risk for schizophrenia. It is not known whether PBC's are associated with increased risk for bipolar disorder and other affective psychoses; whether the increase in risk is explained by life style factors of the mother associated with lower socioeconomic status; whether the increased risk is an expression of genetic predisposition to disorder; whether PBC'S, and/or family history of disorder, are associated with early age of onset; the degree to which intergenerational transmission of socioeconomic status is explained by PBC's, in the general population; and in general, how the major risk factors for severe mental disorders combine in raising risk. Continuation of research on the relationship of PBC's to severe mental disorder (especially schizophrenia, bipolar disorder, autism, and their associated disorders) is proposed. Prior analyses, based on linkage of the medical birth register and psychiatric case register of Denmark for births 1973-1993, showed that PBC's were related prospectively to the entire range of hospitalized mental disorders, in both adults and children, with the possible exception of affective psychosis. This continuation expands: 1) the data collection years through 1999; 2) the case-finding beyond hospitals to include specialty outpatient clinics; and 3) the linkage to include the Central Population Register and the National Bureau of Statistics. A reference sample of 5 percent sample of the population of Denmark, including over 60,000 births after 1973 will be linked as well. The expansion will: 1) generate an expected 1200 cases of schizophrenia, 350 cases of affective psychosis, and 1150 cases of autism; 2) permit linkage of first degree relatives, including siblings, mothers and fathers; and 3) permit linkage to socioeocomic variables, obtained on yearly basis, when taxes are paid, such as occupation, income, education, employment, and wealth. Analytic procedures are mostly in the form of logistic regression, and its extensions to time-to-event models (Proportional Hazards); dependent variables with more than two categories (Polytomous Logistic); and to dependent observations (Generalized Estimating Equations). Denmark's system of social medicine provides free care for all, minimizing leakage of cases; and the system of registers is as highly developed and reliable as anywhere in the world. Results will clarify the etiologies of the severe mental disorders; improve understanding of nosologic relationships between the several disorders; provide quantitative information relevant to planning of longitudinal observational and preventive intervention research; and suggest subsets of cases to improve power for genetic analyses.